Gary Marcus, who among other things is a graduate of Hampshire College and a professor at NYU, has written what I think is a thoughtful and sobering piece on the dangers of recent developments in AI. Discussion on the points he raises here would be most welcome in this thread.